Hindawi Publishing Corporation

Mediators of Inflammation
Volume 2013, Article ID 312476, 12 pages
http://dx.doi.org/10.1155/2013/312476

Research Article
The Causative Pathogen Determines the Inflammatory
Profile in Cerebrospinal Fluid and Outcome in Patients with
Bacterial Meningitis
Denis Grandgirard,1 Rahel Gumann,1 Boubacar Coulibaly,2 Jean-Pierre Dangy,3 Ali Sie,2
Thomas Junghanss,4 Hans Schudel,1 Gerd Pluschke,3 and Stephen L. Leib1,5
1

Neuroinfection Laboratory, Institute for Infectious Diseases, University of Bern, Friedbuehlstrae 51, 3010 Bern, Switzerland
Centre de Recherche en Sante de Nouna, Nouna, Burkina Faso
3
Swiss Tropical and Public Health Institute and University of Basel, 4051 Basel, Switzerland
4
Section of Clinical Tropical Medicine, Heidelberg University Hospital, 69120 Heidelberg, Germany
5
Biology Division, Spiez Laboratory, Federal Office for Civil Protection (FOCP), 3700 Spiez, Switzerland
2

Correspondence should be addressed to Stephen L. Leib; stephen.leib@ifik.unibe.ch

Received 22 February 2013; Revised 28 May 2013; Accepted 4 June 2013
Academic Editor: Jonathan P. Godbout
Copyright 2013 Denis Grandgirard et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.
Background. The brains inflammatory response to the infecting pathogen determines the outcome of bacterial meningitis (BM), for
example, the associated mortality and the extent of brain injury. The inflammatory cascade is initiated by the presence of bacteria in
the cerebrospinal fluid (CSF) activating resident immune cells and leading to the influx of blood derived leukocytes. To elucidate the
pathomechanisms behind the observed difference in outcome between different pathogens, we compared the inflammatory profile
in the CSF of patients with BM caused by Streptococcus pneumonia ( = 14), Neisseria meningitidis ( = 22), and Haemophilus
influenza ( = 9). Methods. CSF inflammatory parameters, including cytokines and chemokines, MMP-9, and nitric oxide synthase
activity, were assessed in a cohort of patients with BM from Burkina Faso. Results. Pneumococcal meningitis was associated with
significantly higher CSF concentrations of IFN-, MCP-1, and the matrix-metalloproteinase (MMP-) 9. In patients with a fatal
outcome, levels of TNF-, IL-1, IL-1RA, IL-6, and TGF- were significantly higher. Conclusion. The signature of pro- and antiinflammatory mediators and the intensity of inflammatory processes in CSF are determined by the bacterial pathogen causing
bacterial meningitis with pneumococcal meningitis being associated with a higher case fatality rate than meningitis caused by N.
meningitidis or H. influenzae.

1. Introduction
The three major pathogens causing bacterial meningitis (BM)
are Streptococcus pneumoniae (SP), Haemophilus influenzae
type b (Hib) and Neisseria meningitidis (NM). BM is the
most severe and frequent infection of the central nervous
system (CNS) and is associated with a high mortality rate and
adverse neurological outcome in a substantial proportion of
survivors [1]. BM caused by SP has the highest case fatality
and neurological disability rates compared to those caused by
NM or Hib [2, 3]. In a recent systematic review, the median
inhospital case fatality ratio among African children with BM
was 35% for SP, 25% for Hib, and 4% for NM meningitis [4]. In

addition, about a quarter of children surviving pneumococcal meningitis and Hib meningitis had neuropsychological
sequelae by the time of hospital discharge.
A number of factors have been identified as predictive
for a poor outcome in terms of mortality. Coma and seizures
were found to be predictive, next to shock, peripheral circulatory failure, severe respiratory distress, a low peripheral white
blood cell (WBC) count, and a high CSF protein level in a
recent systematic review of prognostic studies [5].
The host inflammatory reaction in the CNS is initiated
by the recognition of the invading pathogens and results
in the local production of soluble mediators. Differences in

2
the innate immune responses upon stimulation with Grampositive and Gram-negative bacteria have been demonstrated
in vitro and in experimental infection models. These differences are presumably related to pathogen-specific activation
of pattern recognition receptors [69]. Brain cells, that is,
astrocytes, microglial cells, endothelial cells, ependymal cells,
and resident macrophages, react to the invading pathogens
by releasing early response inflammatory cytokines, like IL1, TNF-, and IL-6. TNF- stimulates the recruitment of
neutrophils and monocytes to the sites of infection and
activates these cells to eliminate pathogens, by releasing
reactive molecules, amongst others NO. Antibiotics causing
rapid lysis of the bacteria have been shown to exacerbate CSF
inflammation by increasing TNF- [10]. After stimulation
by bacterial wall components or TNF-, IL-1 is released
by mononuclear phagocytes, glial cells, and endothelial cells.
High CSF level of IL-1 significantly correlates with adverse
outcome and severity of BM [11]. Administration of TNF-
or IL-1 into the CSF results in pathophysiological changes
characteristic of BM [12, 13]. IL-6 is produced by monocytes,
endothelial cells, and astrocytes, mainly in response to IL-1
[14]. IL-10 and IL-1RA antagonize the effect of proinflammatory cytokines or chemokines, by inhibiting their production
(IL-10) or acting as a decoy receptor (IL-1RA). CSF levels of
other cytokines and chemokines (IL-2, IL-8, IFN-, MCP-1,
MIP-1, and G-CSF) have also been found elevated in BM [15
19]. White blood cells invading the CSF release MMPs and
reactive molecules [2022] which are critically involved in the
pathogenesis of brain damage in BM. Therapeutic strategies
targeting MMPs and oxidative radical have yielded promising
results, albeit limited to experimental BM models [2024]
to date. Tissue-destructive agents released by leukocytes and
brain resident cells, like matrix-metalloproteinases (MMPs)
and oxidants, also mediate brain damage in BM [22]. In BM,
MMPs are involved in the blood-brain barrier opening, in
immune cell extravasation, in the release of cytokines and
cytokine receptors, and in the development of neuronal damage. In patients, elevated CSF levels of MMP-9 and MMP-8
have been detected [21], and high levels of MMP-9 were identified as a risk factor for sequelae [25]. Nitric oxide (NO) has
been shown to contribute to the pathophysiology of meningitis with a phase-dependent role at the level of the cerebral
vasculature by hyperemic effects in early phase and vasodilative effects protecting against ischemia in later phase [26].
Since all of the above-detailed inflammatory mediators
have been shown to influence the outcome in experimental
models of BM, we set out to determine in patients with BM
the association between the CSF concentration of these mediators with the causative organism and the mortality. To this
end the pathogen-specific inflammatory profiles caused by S.
pneumoniae, N. meningitidis, and H. influenzae were analyzed
in CSF from BM patients. The findings from this study may
help understand, at a pathophysiological level, the difference
in outcome observed between the different pathogens.

2. Materials and Methods

2.1. CSF Samples. CSF samples were collected in the Nouna
Health District (NHD), Burkina Faso, during two consecutive

Mediators of Inflammation
meningitis seasons [27]. Ethical clearance for the meningitis
study was obtained from the Comite Local dEthique de
Nouna (Nouna Local Ethical Committee). Procedures followed were in accordance with the ethical standards of the
committee and with the Helsinki Declaration of the World
Medical Association. Informed consent was obtained from
all study participants. Following the national guidelines for
meningitis surveillance, diagnostic lumbar puncture was performed on patients with a suspicion of meningitis presenting
to one of the 25 health centers of the NHD. Patients were
enrolled into the study if their CSF could be transported on
ice and analyzed by trained personnel in the laboratory of
Nouna District Hospital within 6 hours. For primary analysis,
white blood cell counts were determined. Samples were tested
for bacterial pathogens using Gram staining, culture, latex
agglutination, or PCR. CSF samples were centrifuged to
remove white blood cells and supernatants stored at 80 C.
Samples with conflicting diagnostics for the etiological agent
between culture and PCR were excluded from the analysis.
During transport from Africa to Switzerland, samples were
kept frozen in liquid nitrogen.
CSF samples with confirmed acute BM (positive culture and/or positive PCR, CSF WBC of more than 50106
cells/liter) were categorized into three analytical groups
according to the causative agent: S. pneumoniae (SP, = 14),
N. meningitidis (NM, = 22), and H. influenzae type b (Hib,
= 9).
2.2. Assessment of Cytokine Levels in CSF Samples. Cytokine
levels in CSF samples were assessed using microspherebased multiplex assays (Lincoplex, LINCO Research Inc.,
St. Charles, MA, USA). CSF concentrations of the following
cytokines were measured: IL-1, IL-1, IL-2, IL-6, IL-8, IL10, IL-1RA, IFN-, MCP-1, MIP-1, MIP-1, TGF-, and
TNF-. To fit the dynamic range of the test, samples were
assessed undiluted or diluted 5- to 25-fold with the provided
assay buffer, depending on the expected concentration of the
respective analytes to be tested, as determined in preliminary
experiments. A minimum of 50 beads per analyte was
measured. Calibration curves from the provided standards
were calculated using BioPlex Manager software version 4.1.1
with a five-parametric logistic curve fitting. When measured
cytokine concentrations were below the detection limit, a
value corresponding to the detection limit of the assay
multiplied by the dilution factor of the sample was used for
statistical analysis.
Validation of the assay was done for IL-10 and TNF-
using Enzyme-linked immunosorbent assays (ELISA) (R&D
Systems Inc., Minneapolis, MN, USA). According to the
concentrations estimated using the Lincoplex assay and the
sensitivity range of the ELISA, samples were diluted 10fold (TNF-) or 20-fold (IL-10) to a final volume of 200 L
using the appropriate calibrator diluent. Results obtained
by Luminex and conventional ELISA were compared for
correlation, using Prism Software. For both cytokines, a
significant correlation was found between the two methods
(TNF-, = 0.002, Spearman = 0.53; IL-10, < 0.0001,
Spearman = 0.75).

Mediators of Inflammation
2.3. Assessment of MMP-9 Levels in CSF Samples. MMP-9
levels were assessed using the Fluorokine MAP Human MMP
Kit (R&D Systems Inc., Minneapolis, MN, USA). All CSF
samples were diluted 100-fold, to a final volume of 50 L. A
minimum of 50 beads was measured. Standard curves were
calculated similarly to those of the cytokines assay.
Validation of the assay was done using gelatin-containing
gel zymography as already described [21]. Concentrations
measured by the Fluorokine MAP assay correlated with those
assessed by gelatin zymography with a two-tailed -value of
<0.0001 (Spearmans rank correlation test, = 0.66).
2.4. Measurement of Total Nitrate and Nitrite in CSF Samples.
CSF levels of total nitrate and nitrite were assessed using a
nitrate/nitrite colorimetric assay (Cayman Chemical Company, Ann Arbor, MI, USA). The estimated concentrations
were used as an index for nitric oxide synthase activity. CSF
samples were filtered for 30 min at 10000 g using Ultrafree
0.5 centrifugal filter devices. Samples and assay buffer (each
40 L) were mixed with 10 L of coenzyme mixture and
10 L of nitrate reductase in a 96-well plate. After 3 h at
room temperature (RT) for conversion of nitrate to nitrite,
Griess reagents were added for 10 min at RT. Absorbance was
measured at 550 nm. Total nitrite concentrations were calculated using standard curves generated by the SoftMax PRO
software version 3.1.2 (Molecular Devices Inc., Sunnyvale,
CA, USA) using a linear curve fitting.
2.5. Statistical Analysis. Statistical analysis was done using
GraphPad Prism version 5.04 (GraphPad Software Inc., La
Jolla, CA, USA). For comparison of the different pathogen
groups, we first tested whether data sets followed a Gaussian
distribution. At least one group did not follow a Gaussian
distribution for each comparison. Furthermore, since we had
to include arbitrary values, the nonparametric Kruskal-Wallis
test was used. If the overall test was significant ( < 0.05), the
Mann-Whitney test was applied to perform pairwise comparisons. For the analysis of the relation between CSF cytokine
levels and the outcome of the disease, the nonparametric
Mann-Whitney test was used, with a confidence interval of
95% and two-tailed values. Correlations were analyzed
using Spearmans rank correlation test, with a confidence
interval of 95% and two-tailed values.

3. Results
3.1. Clinical Parameters. Significant pathogen-specific differences in the age distribution of patients were observed within
the study cohort ( Kruskal-Wallis test: < 0.01). NM
meningitis was found in patients 060 years (median: 5.5
years, = 22) and SP meningitis in patients 040 years
(median: 5.5 years, = 14). In contrast, Hib meningitis
affected only children 14 years (median: 2 years, = 9). The
difference in median ages was significant between SP versus
Hib ( < 0.05) and NM versus Hib ( < 0.01). Mortality of
BM patients was 46% for SP and 27% for NM, while all nine
patients infected with Hib survived (Table 1).

3
CSF white blood cell counts did not significantly differ within the 3 groups (median SP: 7020 106 /L [100
64000]; median NM: 4900 106 /L [10038560]; median Hib:
5540 106 /L [27220000]) (Table 2).
3.2. Cytokine and Chemokine Levels in CSF. Cytokines
and chemokines showed significantly different regulations
between the causative bacteria (Table 2 and Figure 1). In
particular, the CSF concentration of IFN- was significantly
higher in patients infected with SP compared to NM ( <
0.005) and Hib ( < 0.005). The CSF concentration of IFN-
correlated with the age of the patients ( = 0.15, Pearsons =
0.38; Figure 2). This correlation was significant for meningitis
caused by SP ( = 0.01, = 0.79) and NM ( = 0.02,
= 0.64). This suggests that, during meningitis, adults are
more apt to react with IFN- production than children [28].
Since there was a statistically significant difference in age for
the Hib group, we cannot exclude that the difference in IFN level in this population may be due to the age, rather than
the pathogen. For MCP-1 significant differences between SP
versus NM patients ( = 0.045) and SP versus Hib patients
( < 0.01) were observed (Table 2). In addition, a nonsignificant trend for SP causing higher levels of IL-1 ( < 0.07) and
IL-6 ( = 0.055) was found. Taken together, reciprocal trends
in the association of pro- and anti-inflammatory cytokines
and chemokines with BM caused by the different pathogens
were observed. Il-1, IFN-, and MCP-1, as prototypical
proinflammatory factors, showed higher CSF concentrations
in the patients infected with SP than by NM and Hib. In
contrast, the anti-inflammatory mediators IL-10 and IL-1RA,
were more increased in CSF of patients infected with NM and
Hib. The ratio of pro- to anti-inflammatory mediators, in particular the IL-1/IL-1RA ratio, showed statistically significant
differences, being higher in SP versus NM ( < 0.01) and SP
versus Hib ( < 0.03). Similar correlations were found for IL6/IL-10 and IL-6/IL-1RA ratios (Figure 3). Cyto-/chemokines
concentrations were significantly higher in patients infected
with any of the 3 pathogens when compared with a group
of 7 healthy control patients, as defined by no clinical signs
of meningitis and no increase in WBC in the CSF (median
4 106 cells/L). For IL-1, IL-2, TNF, IFN, and MIP1 , the
majority of these samples were under detection limit, even
when samples were analyzed undiluted (Table 3).
Since the host inflammatory reaction during BM is an
important determinant of disease severity and mortality,
the association between CSF cytokine and chemokine levels
and outcome (survival or death) was investigated (Table 4(a)
and Figure 4). When all patients were analyzed together, a
significant association between fatal outcome and CSF levels
of 5 cytokines, namely, IL-1, TNF-, IL-1RA, IL-6 ( <
0.05), and TGF- ( < 0.02), was found. When pathogens
were investigated separately, TNF- was significantly higher
in patients who died from SP meningitis (Table 4(b)), while
only IL-1RA was significantly higher in patients with a fatal
outcome after NM meningitis (Table 4(c)).
3.3. MMP-9, Nitrate and Nitrite Levels in CSF. In children with bacterial meningitis, matrix-metalloproteinase(MMP-) 9 in the cerebrospinal fluid has been associated with

Mediators of Inflammation

Table 1: Patient groups characteristics.

Pathogen
S. pneumoniae
N. meningitidis
H. influenzae
1
2

Number of patients

Gender
(male/female)

Median age
(minmax)

Mortality (%)
(survivors/death)

14
22
9

5/81
10/92
7/2

5.5 (040)
5.5 (160)
2 (14)

46% (7/6)1
27% (16/6)
0% (9/0)

1 patient not documented.

3 patients not documented.

Table 2: CSF inflammatory parameters of patients infected with SP ( = 14), NM ( = 22), and Hib ( = 9).
Streptococcus pneumoniae
Median (min.max.)
(pg/mL)

Neisseria meningitidis
Median (min.max.)
(pg/mL)

Haemophilus influenzae
Kruskal-Wallis Between group
Median (min.max.)
significance
(pg/mL)

IL-1
IL-2

116.2 (1.791040)
3.28 (0.3819.81)

23.01 (0.61325.9)
1.41 (0.3825.69)

13.3 (0.77112.5)
1.6 (0.183.83)

0.069
ns

(b)

IL-6
IL-10

106232 (4440175553)
7515 (509.544609)

93316 (5917162002)
18604 (1883217931)

23257 (3917112594)
12692 (1466135980)

0.0551
ns

(b)

IL-1RA

103123 (4413983243)

243089 (3183639929)

216530 (77121301000)

ns

TNF-

233.8 (34.851199)

318.2 (17.63390)

170.3 (43.631395)

ns

IFN-
MCP-1
MIP-1

58.99 (8.032853)
10109 (169123567)
829.1 (30.753828)

8.94 (1.77219.8)
1896 (309.840005)
1469 (133.827027)

8.62 (0.4220.03)
2059 (411.43897)
1152 (135.612782)

<0.01
<0.04
ns

MIP-1

3344 (575.87518)

3263 (530.790852)

2563 (136422695)

ns

(a), (b)
(a), (b)

TGF-

73.4 (16.36302.3)

65.26 (10.3146.5)

43.23 (27.52305.2)

ns

MMP-9
WBC

1.51 106 (2756324.5 106 )

7020 (10064000)

525821 (18775.32 106 )

4900 (10038560)

334058 (124402.006 106 )

5540 (27220000)

<0.03
ns

(a), (b)

19.8 (9.5104)

37 (19.24124.1)

21.36 (4.53341.5)

0.059

(c)

Nitrite/nitrate

The nitrite/nitrate concentration (NO) was determined in a subset of CSFs, due to limitations in the available sample volumes (Hib = 8, SP = 8, and NM =
12). The column entitled Between group significance describes statistical significance between groups as determined by Mann-Whitney test, for the following
comparisons (a) SP versus NM, (b) SP versus Hib, and (c) NM versus Hib.

Table 3: CSF inflammatory parameters in control patients ( = 7).

Analytes

Samples under detection limit

Median (pg/mL)
[min.max.]

Dilution factor

Detection limit

IL-1
IL-2
IL-6
IL-10
IL-1RA
TNF-
IFN-
MCP-1
MIP-1
MIP-1
TGF-
MMP-9

5/7
7/7
0/7
1/7
0/7
4/7
7/7
0/7
2/7
4/7
0/7
0/7

n.d
n.d.
115 [95.06175.1]
39.96 [0.41291.3]
184.6 [99.822958]
n.d
n.d
326 [11611355]
35.43 [1.23536.6]
n.d
24.16 [13.7730.3]
106.8 [10.931153]

1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:5
1:1
1:1
1:1
1 : 10

0.19
0.38
0.79
0.41
10.76
0.22
0.55
0.63
1.23
27.65
0.69
n.d.

A median value was calculated only when the majority of samples were above the limit of detection. Control samples were measured undiluted or diluted 1 : 5,
respectively, 1 : 10.
1
Detection limit as provided by the manufacturer.

Mediators of Inflammation

104

105

103

102

MCP-1 (pg/mL)

IFN- (pg/mL)

104

101

103

100

101

SP

NM

102

Hib

SP

NM

(a)

Hib

(b)
10

MMP-9 (pg/mL)

106

105

104

SP

NM

Hib

(c)

Figure 1: Inflammatory CSF parameters in BM patients. Statistically significant differences ( < 0.05;
of MCP-1, IFN-, and MMP-9 were observed in patients with BM grouped for the causative pathogens.

blood-brain barrier damage and neurological sequelae [25].

Concentrations of MMP-9 were highest in CSF of patients
suffering from SP meningitis ( < 0.03; Table 1 and Figure 1).
Pairwise comparisons between the different etiological agents
revealed SP versus Hib to differ significantly ( < 0.02) as
well as SP versus NM ( < 0.04) (Figure 1). MMP-9 showed
a nonsignificant trend towards higher CSF levels in patients
who died from the disease (Mann-Whitney test: = 0.068,
Table 4(a)).
Levels of total nitrate and nitrite showed a nonsignificant
trend ( 0.06) between pathogens, with higher CSF

< 0.01) in CSF concentrations

concentrations in samples of NM patients than in SP and Hib

patients (Table 1).

4. Discussion
In addition to the high mortality of up to 30%, cases of
BM and specifically those caused by SP are associated with
persistent neurological sequelae in up to 50% of the survivors
due to different forms of brain damage [29, 30]. The burden
of disease is especially high in low-income countries, and risk
of mortality or major sequelae is twice as high in African as

Mediators of Inflammation
104

104

103

103

CSF concentration (pg/mL)

CSF concentration (pg/mL)

102

101

102

101

100

100

101

101

10

20

30

40

50

60

70

10

20

30

Age

40

50

Age

SP
NM
Hib

(b)

(a)
10

CSF concentration (pg/mL)

103

102

101

100

101

10

20

30

40

50

60

70

Age
(c)

Figure 2: Correlation between CSF levels of IFN- and age of the patients. (a) The level of IFN- correlated with the age of the patients. This
correlation was significant for SP patients ((b): = 0.01, = 0.79, and black dots) and NM patients ((c): = 0.02, = 0.64, and grey dots).

in the European regions [31]. Over the last four decades, the
risk of major postdischarge sequelae caused by meningitis has
not significantly changed [31]. Both clinical and experimental
studies suggest that both the pathogen and the inflammatory
host response contribute to the development of mortality and
neurological sequelae.

Here we compared the host immune response in the

CSF to BM caused by S. pneumoniae, N. meningitidis, and
H. influenzae. To date, only few studies have compared the
CSF concentration of inflammatory mediators during BM
in relation to the bacterial pathogen [17, 28, 32]. Here we
found that the pathogen is an important determinant of

Mediators of Inflammation

7
IL-6/IL-10

IL 1/IL1RA

0.1

1000

0.01
100

Ratio

Ratio

0.001
10

0.0001

0.00001

0.000001

0.1
SP

NM

Hib

SP

(a)

NM

Hib

(b)
IL-6/IL-1RA

100

Ratio

10

0.1

0.01
SP

NM

Hib

(c)

Figure 3: Ratio of pro- to anti-inflammatory mediators. Statistically significant differences ( < 0.05) in the ratios of IL-1/IL-1RA, IL6/IL-10, and IL-6/IL-1RA were observed in patients with BM grouped for the causative pathogens.

the inflammatory CSF reaction to BM. The observed difference in inflammation in the CSF may not only be due to
inherent differences between pathogens to elicit a response in
cells of the innate immune system [69] but also due to the
ability of the pathogen to multiply in the CNS compartment.
Unfortunately, determining the bacterial load in the CSF of
patients was not feasible in the present study.
In accordance with published data [17, 28], we observed
significantly higher CSF concentration of IFN- in pneumococcal meningitis. Furthermore, as observed by others,
the level of IFN- correlated with the age of the patients.

This suggests that, during meningitis, adults are more apt to

react with IFN- production than children [28]. IFN- is a
potent proinflammatory cytokine. It enhances the function
of macrophages and polymorphonuclear leukocytes by stimulating nonspecific defense mechanisms such as phagocytosis
and the release of inflammatory mediators and may therefore
contribute to the overshooting inflammation.
We found elevated levels of the chemokines MCP-1, MIP1, and MIP-1 in the CSF of patients with BM, in accordance
with other published studies [18, 19, 33]. CSF concentrations
of MCP-1 were significantly higher in patients infected with

Mediators of Inflammation

104

2 105

1.5 105

102

IL-6 (pg/mL)

IL-1 (pg/mL)

103

101

1 105

5 104

100

101

(a)
10

(b)

10

TNF- (pg/mL)

IL-1RA (pg/mL)

106

105

103

102

104

103

101

(c)

(d)

TGF- (pg/mL)

10

102

101

100

(e)

Figure 4: Significant differences in cytokines CSF levels in relation to disease outcome. Patients were grouped independently of the etiological
agents based on the outcome (nonlethal, white boxes/lethal, grey boxes) of the disease. Pairwise comparisons (Mann-Whitney test) revealed
statistically significant differences for the 5 cytokines (see also Table 4).

Mediators of Inflammation

Table 4: Difference in CSF inflammatory parameters between patients with a nonfatal and a fatal outcome.
(a) Pooled pathogens

IL-1
IL-2
IL-6
IL-10
IL-1RA
TNF-
IFN-
MCP-1
MIP-1
MIP-1
TGF-
MMP-9
WBC
Nitrite/nitrate

Nonfatal cases
Median (min.max.) (pg/mL)
13.19 (0.611040)
1.6 (0.18255.69)
67789 (3917162002)
15764 (509.5217931)
117918 (31831.301 106 )
195.1 (17.63390)
11.16 (0.421920)
2101 (309.840005)
935.6 (133.827027)
2728 (530.724661)
41.05 (10.30305.2)
463489 (18775.319 106 )
4600 (100186000)
137.5 (23.15625)

Fatal cases
Median (min.max.) (pg/mL)
57.19 (9.84682)
1.41 (0.386.9)
104035 (77100175553)
14131 (80990755)
337610 (19362983243)
484.3 (88.411199)
49.13 (1.772853)
2903 (118720728)
1166 (30.7516122)
3555 (126390852)
93.3 (36.74302.3)
1.258 106 (3135414.497 106 )
11100 (100042400)
226.9 (83.6368.5)

Mann-Whitney
0.046
0.6618
0.0365
0.9351s
0.0209
0.0436
0.1823
0.2193
0.9351
0.4177
0.0099
0.0671
0.062
0.3627

(b) SP only

IL-1
IL-2
IL-6
IL-10
IL-1RA
TNF-
IFN-
MCP-1
MIP-1
MIP-1
TGF-
MMP-9
WBC

Non-fatal cases ( = 7)
Median (min.max.) (pg/mL)
54.44 (1.791040)
7.87 (0.3819.81)
106232 (4440150456)
7515 (509.544609)
55240 (4413275000)
70.65 (0.22570.9)
40.82 (8.031920)
10109 (169123059)
829.1 (256.93828)
2732 (575.86685)
35.52 (16.36153.5)
1.103 106 (2756324.254 106 )
4600 (10064000)

Fatal cases ( = 5)
Median (min.max.) (pg/mL)
206.9 (9.84682)
1.6 (0.385.55)
136215 (77100175553)
7497 (80917392)
121576 (19362983243)
416.2 (88.411199)
190.7 (23.732853)
2897 (202020307)
416.4 (30.751738)
3771 (12637342)
100.5 (36.74302.3)
2.676 106 (4751124.497 106 )
12930 (320042400)

Mann-Whitney
0.5253
0.1452
0.3434
0.6010
0.1591
0.048
0.202
0.8207
0.5025
0.6313
0.149
0.2331
0.1375

(c) NM only

IL-1
IL-2
IL-6
IL-10
IL-1RA
TNF-
IFN-
MCP-1
MIP-1
MIP-1
TGF-
MMP-9
WBC

Non-fatal cases ( = 15)

Median (min.max.) (pg/mL)
10.56 (0.61325.9)
1.41 (0.3825.69)
67789 (5917162002)
16715 (1883217931)
182057 (3183305690)
267.2 (17.63390)
9.1 (2.62219.8)
1283 (309.840005)
1200 (133.827027)
3137 (530.724661)
44.05 (10.3146.5)
585934 (18775.319 106 )
4900 (10017000)

Fatal cases ( = 5)
Median (min.max.) (pg/mL)
50.5 (10.48205.4)
1.22 (0.386.9)
1100160 (77742145357)
47476 (543690755)
447933 (275000639929)
552.4 (247.41126)
8.3 (1.77102.7)
2909 (118720728)
1937 (585.116122)
3388 (185190852)
86.05 (58.93144.7)
525821 (3135412.188 106 )
7800 (100038560)

Significantly higher CSF concentrations in patients with a fatal outcome are represented in italic lines.

Mann-Whitney
0.3056
0.916
0.3056
0.1974
0.001
0.2661
0.5413
0.3056
0.444
0.5528
0.0526
0.7996
0.3983

10
SP compared to both NM and Hib. Elevated CSF levels
of MCP-1 have been shown to exacerbate brain damage
during neuroinflammatory diseases by increasing the influx
of monocytes and neutrophils [34]. However, we did not find
a correlation of MCP-1 levels with WBCs in the CSF, which
is in accordance with other studies [18, 33]. Furthermore,
WBC count did not correlate with any other parameters,
either when pathogen groups were pooled together or when
analyzed separately. Worsening of the outcome by an overshooting inflammatory reaction is further suggested by the
increased ratios of pro- to anti-inflammatory mediators (IL6/IL-10, IL6/IL-1RA, and IL-1/IL-1RA) observed in patients
infected by SP. In a previous study in BM patients, CSF
showed higher ratio of TNF-/IL-10 by SP when compared
to NM and Hib combined [17].
High CSF concentration of MMP-9 is a risk factor for
a detrimental outcome [25, 35, 36]. Our results add further
support to the notion that MMP-9 is critically involved in the
increase in mortality and sequelae, since CSF levels of MMP-9
were significantly higher in BM caused by SP, which is usually
associated with a higher incidence of neurological sequelae
and mortality [25].
When the relationship between inflammatory mediators
and the outcome of the disease was investigated independently of the causative agent, higher CSF levels of TNF, IL-1, IL-1RA, IL-6, and TGF- were found in patients
who died from BM. A nonsignificant trend was also found
for MMP-9 and IL-6. Thus, the present study confirms
previous observations showing that both a strong activation
of the IL-1 system [37] and increased MMP-9 levels [25]
correlate with adverse outcome of BM. While the higher
mortality observed in SP may be seen in the context of
higher cytokine levels, this observation could not be made
for the difference in mortality between NM (27%) and
Hib (0%) where inflammatory CSF parameters were not
significantly different. In comparison to other studies in
which mortality in developing regions reached 30% [38],
the mortality attributed to Hib was exceptionally low in the
present study. The small group size for Hib is a clear limitation
of the study. Furthermore, NM meningitis is more often
associated with fulminant septicemia, which may contribute
to mortality. Unfortunately, data concerning the presence of
concomitant septicemia were not available in the present
study. Interestingly, the inhibition of the metalloproteinase
TACE/ADAM17, acting as a sheddase for TNF- and TGF [39], has been shown to lower mortality and to attenuate
brain injury in experimental models of BM [20, 36]. In line
with these results, we could show in the present study that
TNF- was significantly upregulated in patients with a poor
outcome. In accordance with this clinical observation is the
experimental finding that deletion of another member of the
TGF family, TGF- has been shown to improve bacterial
clearance and diminished intracranial complication in a
mouse with pneumococcal meningitis [40].
A nonsignificant trend for higher levels of nitrite/nitrate
levels was observed in NM infected patients. In experimental
models, CSF NO/nitrite concentration correlated with an
increase in blood-brain barrier permeability, but inhibition
of the different nitric oxide synthases resulted in inconsistent

Mediators of Inflammation
effects, probably as a result of differences in the timing
of intervention and the corresponding effects on the brain
perfusion.
The present study, analyzing a cohort of patients affected
by meningitis, identified several factors which contribute
to the worsening of outcome in bacterial meningitis. Interestingly, some of these factors (TNF-, MMPs, and nitric
oxide) have already been described in experimental models
using knockout animals [41] and/or intervention strategies
[42] which reduced mortality and ameliorated the outcome
of infected animals. The most promising strategies derived
from these experimental models include the reduction of
the inflammatory reaction by targeting different steps in the
inflammatory process [43], from the release of proinflammatory bacterial products to the activation of the innate
immune system and the production/release of cytokines or
chemokines, as well as the inhibition of metalloproteinases
or treatments with antioxidants [44].

5. Conclusion
In conclusion, this study showed that SP, NM, and Hib
elicit distinct profiles of inflammatory mediators in the
CSF during BM. A more intense inflammatory reaction, in
particular higher CSF levels of IFN-, MCP-1, and MMP-9,
were observed in patients infected with SP. Furthermore, the
ratios of pro- to anti-inflammatory parameters were found to
be significantly higher in patients with SP meningitis. This
is likely to contribute to the higher case fatality rate and
morbidity observed in patients suffering from pneumococcal
meningitis and may therefore help find new treatment strategies aimed at improving the outcome of infected patients.

Acknowledgments
The authors thank Franziska Simon, Neuroinfection Laboratory, Institute for Infectious Disease, University of Bern,
Bern, Switzerland, for excellent technical assistance. The
authors report no conflict of interests with the trademarks
and companies mentioned in the present study. This study
was financially supported by the Swiss National Science
Foundation (nos. 116257 and 138094).

References
[1] K. Grimwood, V. A. Anderson, L. Bond et al., Adverse
outcomes of bacterial meningitis in school-age survivors,
Pediatrics, vol. 95, no. 5, pp. 646656, 1995.
[2] D. van de Beek, J. de Gans, L. Spanjaard, M. Weisfelt, J. B.
Reitsma, and M. Vermeulen, Clinical features and prognostic
factors in adults with bacterial meningitis, The New England
Journal of Medicine, vol. 351, no. 18, pp. 18491959, 2004.
[3] A. Schuchat, K. Robinson, J. D. Wenger et al., Bacterial
meningitis in the United States in 1995. Active Surveillance
Team, The New England Journal of Medicine, vol. 337, no. 14,
pp. 970976, 1997.
[4] M. Ramakrishnan, A. J. Ulland, L. C. Steinhardt, J. C. Mosi, F.
Were, and O. S. Levine, Sequelae due to bacterial meningitis

Mediators of Inflammation

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

among African children: a systematic literature review, BMC

Medicine, vol. 7, article 47, 2009.
R. C. J. de Jonge, A. M. van Furth, M. Wassenaar, R. J. B.
J. Gemke, and C. B. Terwee, Predicting sequelae and death
after bacterial meningitis in childhood: a systematic review of
prognostic studies, BMC Infectious Diseases, vol. 10, article 232,
2010.
A. Diab, J. Zhu, L. Lindquist, B. Wretlind, M. Bakhiet, and H.
Link, Haemophilus influenzae and Streptococcus pneumoniae
induce different intracerebral mRNA cytokine patterns during
the course of experimental bacterial meningitis, Clinical and
Experimental Immunology, vol. 109, no. 2, pp. 233241, 1997.
M. I. Fowler, R. O. Weller, J. E. Heckels, and M. Christodoulides,
Different meningitis-causing bacteria induce distinct inflammatory responses on interaction with cells of the human
meninges, Cellular Microbiology, vol. 6, no. 6, pp. 555567,
2004.
T. H. Mogensen, S. R. Paludan, M. Kilian, and L. stergaard,
Live Streptococcus pneumoniae, Haemophilus influenzae, and
Neisseria meningitidis activate the inflammatory response
through Toll-like receptors 2, 4, and 9 in species-specific
patterns, Journal of Leukocyte Biology, vol. 80, no. 2, pp. 267
277, 2006.
K. Tietze, A. Dalpke, S. Morath, R. Mutters, K. Heeg, and C.
Nonnenmacher, Differences in innate immune responses upon
stimulation with gram-positive and gram-negative bacteria,
Journal of Periodontal Research, vol. 41, no. 5, pp. 447454, 2006.
M. M. Mustafa, O. Ramillo, X. Saez-Llorens, K. D. Olsen, R.
R. Magness, and G. H. McCracken Jr., Cerebrospinal fluid
prostaglandins, interleukin 1, and tumor necrosis factor in
bacterial meningitis. Clinical and laboratory correlations in
placebo-treated and dexamethasone-treated patients, American Journal of Diseases of Children, vol. 144, no. 8, pp. 883887,
1990.
M. M. Mustafa, M. H. Lebel, O. Ramilo et al., Correlation of
interleukin-1 and cachectin concentrations in cerebrospinal
fluid and outcome from bacterial meningitis, Journal of Pediatrics, vol. 115, no. 2, pp. 208213, 1989.
O. Ramilo, X. Saez-Llorens, J. Mertsola et al., Tumor necrosis
factor /cachectin and interleukin 1 initiate meningeal inflammation, Journal of Experimental Medicine, vol. 172, no. 2, pp.
497507, 1990.
G. A. Rosenberg, E. Y. Estrada, J. E. Dencoff, and W. G. StetlerStevenson, Tumor necrosis factor--induced gelatinase B
causes delayed opening of the blood-brain barrier: an expanded
therapeutic window, Brain Research, vol. 703, no. 1-2, pp. 151
155, 1995.
F. Rusconi, F. Parizzi, L. Garlaschi et al., Interleukin 6 activity
in infants and children with bacterial meningitis, Pediatric
Infectious Disease Journal, vol. 10, no. 2, pp. 117121, 1991.
S. J. Hackett, A. P. J. Thomson, and C. A. Hart, Cytokines,
chemokines and other effector molecules involved in meningococcal disease, Journal of Medical Microbiology, vol. 50, no. 10,
pp. 847859, 2001.
S. Kastenbauer, B. Angele, B. Sporer, H. Pfister, and U. Koedel,
Patterns of protein expression in infectious meningitis: a
cerebrospinal fluid protein array analysis, Journal of Neuroimmunology, vol. 164, no. 1-2, pp. 134139, 2005.
R. F. Kornelisse, C. E. Hack, H. F. J. Savelkoul et al., Intrathecal
production of interleukin-12 and gamma interferon in patients
with bacterial meningitis, Infection and Immunity, vol. 65, no.
3, pp. 877881, 1997.

11
[18] C. M. Mastroianni, L. Lancella, F. Mengoni et al., Chemokine
profiles in the cerebrospinal fluid (CSF) during the course of
pyogenic and tuberculous meningitis, Clinical and Experimental Immunology, vol. 114, no. 2, pp. 210214, 1998.
[19] H. Sprenger, A. Rosler, P. Tonn, H. J. Braune, G. Huffmann, and
D. Gemsa, Chemokines in the cerebrospinal fluid of patients
with meningitis, Clinical Immunology and Immunopathology,
vol. 80, no. 2, pp. 155161, 1996.
[20] S. L. Leib, J. M. Clements, R. L. P. Lindberg et al., Inhibition of matrix metalloproteinases and tumour necrosis factor
converting enzyme as adjuvant therapy in pneumococcal
meningitis, Brain, vol. 124, part 9, pp. 17341742, 2001.
[21] S. L. Leib, D. Leppert, J. Clements, and M. G. Tauber, Matrix
metalloproteinases contribute to brain damage in experimental
pneumococcal meningitis, Infection and Immunity, vol. 68, no.
2, pp. 615620, 2000.
[22] D. N. Meli, S. Christen, and S. L. Leib, Matrix metalloproteinase-9 in pneumococcal meningitis: activation via an oxidative
pathway, Journal of Infectious Diseases, vol. 187, no. 9, pp. 1411
1415, 2003.
[23] M. Auer, L. Pfister, D. Leppert, M. G. Tauber, and S. L.
Leib, Effects of clinically used antioxidants in experimental
pneumococcal meningitis, Journal of Infectious Diseases, vol.
182, no. 1, pp. 347350, 2000.
[24] S. L. Leib, Y. S. Kim, L. L. Chow, R. A. Sheldon, and M. G.
Tauber, Reactive oxygen intermediates contribute to necrotic
and apoptotic neuronal injury in an infant rat model of bacterial
meningitis due to group B streptococci, The Journal of Clinical
Investigation, vol. 98, no. 11, pp. 26322639, 1996.
[25] D. Leppert, S. L. Leib, C. Grygar, K. M. Miller, U. B. Schaad,
and G. A. Hollander, Matrix metalloproteinase (MMP)-8 and
MMP-9 in cerebrospinal fluid during bacterial meningitis:
association with blood-brain barrier damage and neurological
sequelae, Clinical Infectious Diseases, vol. 31, no. 1, pp. 8084,
2000.
[26] S. L. Leib, Y. S. Km, S. M. Black, J. H. Tureen, and M. G.
Tauber, Inducible nitric oxide synthase and the effect of
aminoguanidine in experimental neonatal meningitis, Journal
of Infectious Diseases, vol. 177, no. 3, pp. 692700, 1998.
[27] A. Sie, V. Pfluger, B. Coulibaly et al., ST2859 serogroup a
meningococcal meningitis outbreak in Nouna Health District,
Burkina Faso: a prospective study, Tropical Medicine and
International Health, vol. 13, no. 6, pp. 861868, 2008.
[28] M. Glimaker, P. Olcen, and B. Andersson, Interferon- in
cerebrospinal fluid from patients with viral and bacterial
meningitis, Scandinavian Journal of Infectious Diseases, vol. 26,
no. 2, pp. 141147, 1994.
[29] K. L. OBrien, L. J. Wolfson, J. P. Watt et al., Burden of disease
caused by Streptococcus pneumoniae in children younger than 5
years: global estimates, The Lancet, vol. 374, no. 9693, pp. 893
902, 2009.
[30] M. Weisfelt, J. de Gans, T. van der Poll, and D. van de
Beek, Pneumococcal meningitis in adults: new approaches to
management and prevention, The Lancet Neurology, vol. 5, no.
4, pp. 332342, 2006.
[31] K. Edmond, A. Clark, V. S. Korczak, C. Sanderson, U. K.
Griffiths, and I. Rudan, Global and regional risk of disabling
sequelae from bacterial meningitis: a systematic review and
meta-analysis, The Lancet Infectious Diseases, vol. 10, no. 5, pp.
317328, 2010.
[32] M. Glimaker, P. Kragsbjerg, M. Forsgren, and P. Olcen, Tumor
necrosis factor- (TNF) in cerebrospinal fluid from patients

12

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

Mediators of Inflammation
with meningitis of different etiologies: high levels of TNF
indicate bacterial meningitis, Journal of Infectious Diseases, vol.
167, no. 4, pp. 882889, 1993.
K. Spanaus, D. Nadal, H. Pfister et al., C-X-C and C-C
chemokines are expressed in the cerebrospinal fluid in bacterial
meningitis and mediate chemotactic activity on peripheral
blood-derived polymorphonuclear and mononuclear cells in
vitro, Journal of Immunology, vol. 158, no. 4, pp. 19561964,
1997.
G. Conductier, N. Blondeau, A. Guyon, J. Nahon, and C. Rov`ere,
The role of monocyte chemoattractant protein MCP1/CCL2 in
neuroinflammatory diseases, Journal of Neuroimmunology, vol.
224, no. 1-2, pp. 93100, 2010.
J. Sellner and S. L. Leib, In bacterial meningitis cortical brain
damage is associated with changes in parenchymal MMP9/TIMP-1 ratio and increased collagen type IV degradation,
Neurobiology of Disease, vol. 21, no. 3, pp. 647656, 2006.
D. N. Meli, J. M. Loeffler, P. Baumann et al., In pneumococcal
meningitis a novel water-soluble inhibitor of matrix metalloproteinases and TNF- converting enzyme attenuates seizures
and injury of the cerebral cortex, Journal of Neuroimmunology,
vol. 151, no. 1-2, pp. 611, 2004.
A. M. van Furth, J. J. Roord, and R. van Furth, Roles of
proinflammatory and anti-inflammatory cytokines in pathophysiology of bacterial meningitis and effect of adjunctive
therapy, Infection and Immunity, vol. 64, no. 12, pp. 48834890,
1996.
H. Peltola, Worldwide Haemophilus influenzae type b disease
at the beginning of the 21st century: global analysis of the
disease burden 25 years after the use of the polysaccharide
vaccine and a decade after the advent of conjugates, Clinical
Microbiology Reviews, vol. 13, no. 2, pp. 302317, 2000.
D. R. Goddard, R. A. D. Bunning, and M. N. Woodroofe,
Astrocyte and endothelial cell expression of ADAM 17 (TACE)
in adult human CNS, Glia, vol. 34, no. 4, pp. 267271, 2001.
U. Malipiero, U. Koedel, W. Pfister, and A. Fontana, Bacterial
meningitis: the role of transforming growth factor-beta in
innate immunity and secondary brain damage, Neurodegenerative Diseases, vol. 4, no. 1, pp. 4350, 2007.
R. Paul, U. Koedel, and H. Pfister, Development of adjunctive
therapies for bacterial meningitis and lessons from knockout
mice, Neurocritical Care, vol. 2, no. 3, pp. 313324, 2005.
B. Woehrl, M. Klein, D. Grandgirard, U. Koedel, and S. Leib,
Bacterial meningitis: current therapy and possible future
treatment options, Expert Review of Anti-Infective Therapy, vol.
9, no. 11, pp. 10531065, 2011.
M. van der Flier, S. P. M. Geelen, J. L. L. Kimpen, I. M. Hoepelman, and E. I. Tuomanen, Reprogramming the host response
in bacterial meningitis: how best to improve outcome? Clinical
Microbiology Reviews, vol. 16, no. 3, pp. 415429, 2003.
D. Grandgirard and S. L. Leib, Meningitis in neonates: bench
to bedside, Clinics in Perinatology, vol. 37, no. 3, pp. 655676,
2010.